Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,65 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
import torch
|
3 |
+
import torchaudio
|
4 |
+
from tortoise.api import TextToSpeech
|
5 |
+
from tortoise.utils.audio import load_voice
|
6 |
+
|
7 |
+
# Initialize Tortoise-TTS
|
8 |
+
tts = TextToSpeech()
|
9 |
+
|
10 |
+
# Function to generate speech
|
11 |
+
def generate_speech(text, voice):
|
12 |
+
# Load the selected voice
|
13 |
+
voice_samples, conditioning_latents = load_voice(voice)
|
14 |
+
|
15 |
+
# Generate speech
|
16 |
+
with torch.no_grad():
|
17 |
+
gen = tts.tts_with_preset(text, voice_samples=voice_samples, conditioning_latents=conditioning_latents, preset="fast")
|
18 |
+
torchaudio.save("output.wav", gen.squeeze(0).cpu(), 24000)
|
19 |
+
|
20 |
+
return "output.wav"
|
21 |
+
|
22 |
+
# Gradio interface
|
23 |
+
def tts_app(text, voice):
|
24 |
+
if not text:
|
25 |
+
return "Please enter some text."
|
26 |
+
return generate_speech(text, voice)
|
27 |
+
|
28 |
+
# Gradio UI
|
29 |
+
with gr.Blocks(
|
30 |
+
theme=gr.themes.Soft(primary_hue="teal", secondary_hue="pink"),
|
31 |
+
css=".gradio-container {background: linear-gradient(135deg, #f5f7fa, #c3cfe2);} "
|
32 |
+
"button {background: linear-gradient(135deg, #6a11cb, #2575fc); color: white; border: none; padding: 10px 20px; border-radius: 5px;} "
|
33 |
+
"button:hover {background: linear-gradient(135deg, #2575fc, #6a11cb);} "
|
34 |
+
) as demo:
|
35 |
+
gr.Markdown(
|
36 |
+
"""
|
37 |
+
# 🎤 **Voice Cloning and Text-to-Speech**
|
38 |
+
**Convert text into natural-sounding speech or clone a voice!**
|
39 |
+
"""
|
40 |
+
)
|
41 |
+
|
42 |
+
with gr.Row():
|
43 |
+
text_input = gr.Textbox(
|
44 |
+
lines=5,
|
45 |
+
label="📝 **Enter Text**",
|
46 |
+
placeholder="Type your text here..."
|
47 |
+
)
|
48 |
+
voice_input = gr.Dropdown(
|
49 |
+
choices=["random", "custom_voice"], # Add more voices as needed
|
50 |
+
label="🗣️ **Select Voice**",
|
51 |
+
value="random"
|
52 |
+
)
|
53 |
+
|
54 |
+
output_audio = gr.Audio(label="🎧 **Generated Speech**", type="filepath")
|
55 |
+
|
56 |
+
submit_button = gr.Button("✨ **Generate Speech**")
|
57 |
+
|
58 |
+
submit_button.click(
|
59 |
+
tts_app,
|
60 |
+
inputs=[text_input, voice_input],
|
61 |
+
outputs=output_audio
|
62 |
+
)
|
63 |
+
|
64 |
+
# Launch the app
|
65 |
+
demo.launch()
|